• AVR Freaks

Helpful ReplyHot!How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup?

Page: 12 > Showing page 1 of 2
Author
Javier Lopez
Pulsotron designer
  • Total Posts : 59
  • Reward points : 0
  • Joined: 2019/02/27 09:59:32
  • Location: Spain
  • Status: offline
2020/01/10 01:32:27 (permalink)
0

How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup?

I have to drive a 3 phase bridge. The problem is when programming the microcontroller it makes all pwm lines to 3.3V so high and low side IGBTs will be ON at the same time shortcircuiting the 800 amps battery to GND.
I have been looking for IGBT bridges that are protected again shortcircuiting both sides at the same time but works well for 1 phase because in a 3 phase I have to set a different sequence. 
post edited by Javier Lopez - 2020/01/10 02:08:03
#1
du00000001
Just Some Member
  • Total Posts : 3315
  • Reward points : 0
  • Joined: 2016/05/03 13:52:42
  • Location: Germany
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/10 02:56:02 (permalink)
0
I doubt the microcontroller "makes all ... high": following reset, all I/O pins are expected to be tri-stated.
 
Anyway: might be best to gate (AND) your control lines with the output from some external watchdog chip. This way, you'd cover most situations (unfortunately not all) that might ruin your power stage.
Other schemes (fuses, HV switch-off, control signal "plausibility checker" etc.) apply as well...

PEBKAC / EBKAC / POBCAK / PICNIC (eventually see en.wikipedia.org)
#2
JPortici
Super Member
  • Total Posts : 879
  • Reward points : 0
  • Joined: 2012/11/17 06:27:45
  • Location: Grappaland
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/10 03:00:50 (permalink)
4.5 (2)
Someone correct me, but the default state of pins during reset (and also during programming) should be input, high impedance, so they are being pulled high from other parts of the circuit, and you should have designed the circuit considering that, expecially at those energy levels. Redesign the board with more attention to safety
#3
Javier Lopez
Pulsotron designer
  • Total Posts : 59
  • Reward points : 0
  • Joined: 2019/02/27 09:59:32
  • Location: Spain
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/10 03:49:33 (permalink)
0
I can place external pull down resistors, but I am still scared about place a monster battery pack connected to a microchip+driver+high power bridge+high power motor without being completely sure if is safe in all situations because it is a survival question
 
A 300 amps fuse is a metal bar...
post edited by Javier Lopez - 2020/01/10 03:50:43
#4
Howard Long
Super Member
  • Total Posts : 744
  • Reward points : 0
  • Joined: 2005/04/04 08:50:32
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/10 04:03:20 (permalink)
0
Jack_M
Someone correct me, but the default state of pins during reset (and also during programming) should be input, high impedance, so they are being pulled high from other parts of the circuit, and you should have designed the circuit considering that, expecially at those energy levels. Redesign the board with more attention to safety




As far as I am concerned, you're right. Irrespective, if I was risking blowing up a $400 downstream device in a system test (with a big bang), I'd be making sure in fairly exhaustive unit tests first that the behaviour is as required. This particularly applies during power up and power down. Furthermore, I would think that separately designed fault conditions should be incorporated at an early stage of development. Almost certainly devising a test jig for firmware development is a reasonable recommendation to avoid $400 of damage when something's not quite right.
#5
JPortici
Super Member
  • Total Posts : 879
  • Reward points : 0
  • Joined: 2012/11/17 06:27:45
  • Location: Grappaland
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/10 04:11:13 (permalink)
0
jlopez2022@gmail.com
I can place external pull down resistors, but I am still scared



As you should be.
This is a scenario where functional safety comes in.
Go back to the drawing board and redesign the PCB but first make a list of every single possible thing you can think of that can go wrong (and ask other people for other inputs), make an estimate of the possibility of the fault happening and find a way to minimize the risk to an acceptable level (i.e.: almost zero chance and/or mitigate the consequences)
#6
du00000001
Just Some Member
  • Total Posts : 3315
  • Reward points : 0
  • Joined: 2016/05/03 13:52:42
  • Location: Germany
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/10 04:26:10 (permalink)
5 (1)
Javier Lopez
A 300 amps fuse is a metal bar...

 
A 300 A fuse is a slightly wider thin sheet of metal (had to repeatedly replace the glow plug fuses on some car - these were rated 50 A). It's just the 300 A conductor that is a bar. But that one isn't expected to significantly warm-up during operation...
 
One might even consider to implement monoflops on the high-side control lines. Won't hinder operation noticeably, but would allow for virtually unlimited debugging  wink

PEBKAC / EBKAC / POBCAK / PICNIC (eventually see en.wikipedia.org)
#7
Javier Lopez
Pulsotron designer
  • Total Posts : 59
  • Reward points : 0
  • Joined: 2019/02/27 09:59:32
  • Location: Spain
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/10 04:29:00 (permalink)
0
Howard Long  I'd be making sure in fairly exhaustive unit tests first that the behaviour is as required.

 
I think it is not enough. To make a safe system it is needed also design it taking in account change of devices changes between versions, specially with microcontrollers where new version arrives with improvements, so I should like know if PWM drivers cames with some protections during startup and programming by default.
 
It would be OK for me if the devices goes to be input until they are programmed because with simple pull down resistors it would be safe. It would be better have the weak pull down activated by default also, specially PWM lines
 
The monoflop would be a good idea to be connected to the enable pin of the "PCB isolated part" power supply
 
post edited by Javier Lopez - 2020/01/10 04:35:35
#8
ric
Super Member
  • Total Posts : 25244
  • Reward points : 0
  • Joined: 2003/11/07 12:41:26
  • Location: Australia, Melbourne
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/10 04:45:35 (permalink)
0
jlopez2022@gmail.com
It would be OK for me if the devices goes to be input until they are programmed because with simple pull down resistors it would be safe.

That is how they always work

It would be better have the weak pull down activated by default also, specially PWM lines

That would only happen if you found a PIC with that option in the config words.
I'm not aware of one.
 

I also post at: PicForum
Links to useful PIC information: http://picforum.ric323.co...opic.php?f=59&t=15
NEW USERS: Posting images, links and code - workaround for restrictions.
To get a useful answer, always state which PIC you are using!
#9
Howard Long
Super Member
  • Total Posts : 744
  • Reward points : 0
  • Joined: 2005/04/04 08:50:32
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/10 05:00:07 (permalink)
0
Is there also a risk at power up, before Vdd reaches a certain level, that the GPIOs are in an indeterminate low impedance state?
#10
Javier Lopez
Pulsotron designer
  • Total Posts : 59
  • Reward points : 0
  • Joined: 2019/02/27 09:59:32
  • Location: Spain
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/10 06:01:02 (permalink)
0
Howard Long
Is there also a risk at power up, before Vdd reaches a certain level, that the GPIOs are in an indeterminate low impedance state?



There are drivers that have UVLO protection. 
I think the best idea is to use half bridge driver that are protected again simultaneous activation of high and low sides,and also as said Howard, with the UVLO protection and isolated.
I found the Si82393, UCC21540 and UCC21530 but are more difficult to find than the risky ones
 
The isolation of the driver is a good idea for avoiding high voltage introduced at battery GND line when passing hundreds amps
post edited by Javier Lopez - 2020/01/10 06:03:36
#11
du00000001
Just Some Member
  • Total Posts : 3315
  • Reward points : 0
  • Joined: 2016/05/03 13:52:42
  • Location: Germany
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/10 06:12:24 (permalink) ☄ Helpfulby Javier Lopez 2020/01/13 01:00:46
1 (1)
The UCC217xx series might be worth a look: while these are single-channel drivers, there are fancy schemes to prevent a shoot-through.

PEBKAC / EBKAC / POBCAK / PICNIC (eventually see en.wikipedia.org)
#12
T Yorky
Super (Thick) Member
  • Total Posts : 537
  • Reward points : 0
  • Joined: 2012/08/28 02:07:35
  • Location: UK
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/10 06:31:36 (permalink)
5 (2)
The pins being input when programming is a presumption. Check your data sheet.
A good example is the programming pins (normally 3 sets on a chip). These all operate in tandem when programming so depends what you have connected to the pins.
Not used your chip you, but definitely MC devices for use with 3 phase modules. I think you need to visit your test procedures. I would NEVER program with the DC bus live. If you using the chip debug mode, again DC bus is never live. Crappy MPLabX resets a PICKIT numerous times when programming how do you predict what will happen in revised firmware under these conditions.
T Yorky.
 
#13
Antipodean
Super Member
  • Total Posts : 1790
  • Reward points : 0
  • Joined: 2008/12/09 10:19:08
  • Location: Didcot, United Kingdom
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/10 09:47:25 (permalink)
5 (3)
Why are people talking about the pin state during programming?
The board with the microcontroller should be programmed and tested before it is connected to anything it could possibly destroy. If it ever needs a software upgrade this should not happen in situ, that is too dangerous, that functionality becomes a swapout to reprogramme it.
Modular construction and replaceable units are the order of the day here.

Do not use my alias in your message body when replying, your message will disappear ...

Alan
#14
du00000001
Just Some Member
  • Total Posts : 3315
  • Reward points : 0
  • Joined: 2016/05/03 13:52:42
  • Location: Germany
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/10 14:10:26 (permalink)
5 (1)
Anti po dean
Modular construction and replaceable units are the order of the day ...

 
By no way! Consider automotive applications where in-system update is the current trend/state of the art.
With some luck you'll have a battery management system that dis connects the battery during updates. But even this is not always the case ...

PEBKAC / EBKAC / POBCAK / PICNIC (eventually see en.wikipedia.org)
#15
Javier Lopez
Pulsotron designer
  • Total Posts : 59
  • Reward points : 0
  • Joined: 2019/02/27 09:59:32
  • Location: Spain
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/13 01:03:05 (permalink)
0
du00000001
The UCC217xx series might be worth a look: while these are single-channel drivers, there are fancy schemes to prevent a shoot-through.

The UCC217xx not are half bridge drivers, so they have not implemented any protection system to avoid both parts of a half bridge to be switched on at the same time



#16
Javier Lopez
Pulsotron designer
  • Total Posts : 59
  • Reward points : 0
  • Joined: 2019/02/27 09:59:32
  • Location: Spain
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/13 01:11:13 (permalink)
0
T Yorky
 I would NEVER program with the DC bus live.

Of course, but disconnect the 300A bus is not easy and somebody can forget that, so the development environment must be designed to be impossible to switch on the both sides at the same time of every half bridge
#17
du00000001
Just Some Member
  • Total Posts : 3315
  • Reward points : 0
  • Joined: 2016/05/03 13:52:42
  • Location: Germany
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/13 03:07:13 (permalink)
0

The UCC217xx not are half bridge drivers ...

 
At the first glance this might not be able to prevent shoot-through. But these chips sport dual inputs (IN+, IN-) that are intended to prevent shoot-through with these single-channel chips. Have a look into the datasheet (e.g. UCC21710, P.35).
 
Anyway, I recalled wrongly: originally I wanted to recommend the UCC23513 - a somewhat simpler variant of the UCC217xx family. A single-channel driver as well - with a quite clever way to prevent shoot-through.

PEBKAC / EBKAC / POBCAK / PICNIC (eventually see en.wikipedia.org)
#18
du00000001
Just Some Member
  • Total Posts : 3315
  • Reward points : 0
  • Joined: 2016/05/03 13:52:42
  • Location: Germany
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/13 03:13:08 (permalink)
0
Javier Lopez
..., so the development environment must be designed to be impossible to switch on the both sides at the same time of every half bridge

 
Just being able to switch one branch to Bus+ and one branch to Bus- will most likely result in overloading, thus destroying the motor (and most likely the respective transistors as well): the bridges are usually designed considering the motor's inductance. Once you drive them statically (as might be the case during updating or debugging), the inductance is virtually zero. Thus the currents are way higher than during normal (dynamic) operation.

PEBKAC / EBKAC / POBCAK / PICNIC (eventually see en.wikipedia.org)
#19
Javier Lopez
Pulsotron designer
  • Total Posts : 59
  • Reward points : 0
  • Joined: 2019/02/27 09:59:32
  • Location: Spain
  • Status: offline
Re: How avoid dspic33EP512GM710 blow batteries and the $400 3 phase bridge at prog/startup 2020/01/13 03:54:05 (permalink)
0

To driver a high capacity IGBT you need a push-pull, so as you can see UCC21710  specs, in+ and in- are linked to only one of the six IGBTs, (and it have a lot of legs to route) so it does not prevent both switchings.
 
Motors windings can be driven during short time a winding continuously, but can make a power down that resets continuously the microcontroller and create other faults (it happened to me :( )
 
It would be good to have an only one isolated driver for a three phase motor, but I did not find any.
 
post edited by Javier Lopez - 2020/01/13 04:28:10
#20
Page: 12 > Showing page 1 of 2
Jump to:
© 2020 APG vNext Commercial Version 4.5